Spatial Analysis
Spatio-Temporal Analysis of Walmart Dataset
This analysis leverages geospatial and temporal data to uncover patterns, trends, and insights related to sales, customer behavior, and operational efficiency in the Walmart dataset.
Key Features Analyzed
- Shipping Duration
- Average Rating
- Order Quantity
- Geographic Distribution
- Temporal Patterns
Analysis Components
- Spatial Data Visualization
- Temporal Trend Analysis
- Customer Behavior Patterns
- Operational Efficiency Metrics
Density Heatmap
Heatmap showing the density of orders across different regions, highlighting areas with high order volumes and potential market opportunities.
Cluster Map
Spatial clustering analysis identifying key market areas and potential expansion opportunities.
Analysis Insights
Spatial Insights
- Identification of high-sales regions
- Underperforming areas analysis
- Geographic distribution patterns
Temporal Insights
- Sales trends over time
- Seasonal patterns
- Peak performance periods
Operational Insights
- Shipping efficiency analysis
- Customer satisfaction correlation
- Regional performance metrics
Cluster Analysis
Cluster 0.0:
Orders: 879 (10.4% of total)
Center: (-31.3576, -34.9008)
States: AE, AP, AZ, CA, CO, GU, ID, KS, MT, ND, NE, NM, NV, OK, PR, SD, TX, UT, VI, WY
Top Cities:
ALBUQUERQUE, APO, Aberdeen, Abernathy, Alamogordo
Cluster 1.0:
Orders: 3771 (44.5% of total)
Center: (-18.9563, -20.8636)
States: AE, AL, CT, DC, DE, FL, GA, IN, KY, MA, MD, ME, MI, NC, NH, NJ, NY, OH, PA, PR, RI, SC, TN, VA, VI, VT, WV
Top Cities:
ALLONS, ARDSLEY, Aaronsburg, Abbottstown, Abington
Cluster 2.0:
Orders: 2493 (29.5% of total)
Center: (-27.4613, -33.0571)
States: AE, AL, AR, CA, FL, IA, IL, IN, KS, KY, LA, MI, MN, MO, MS, ND, NE, OK, SD, TN, TX, WI
Top Cities:
APO, APPLETON, Abbeville, Aberdeen, Abilene
Cluster 3.0:
Orders: 11 (0.1% of total)
Center: (79.1384, 79.1156)
States: GU, MP
Top Cities:
Dededo, Hagatna, Tamuning, Tinian, Yigo
Cluster 4.0:
Orders: 1311 (15.5% of total)
Center: (-42.0634, -40.0731)
States: AK, CA, HI, ID, MT, NV, OR, WA
Top Cities:
Adair Village, Adelanto, Aguanga, Aiea, Algona
Density Analysis
Top 5 States by Order Density:
CA: 869 orders (10.3% of total)
Top cities: Los Angeles (29), San Diego (28), San Jose (18), Bakersfield (17), San Francisco (15)
TX: 695 orders (8.2% of total)
Top cities: Houston (47), Dallas (33), San Antonio (23), El Paso (20), Austin (18)
FL: 663 orders (7.8% of total)
Top cities: Jacksonville (30), Miami (26), Orlando (19), Naples (18), Tampa (15)
NY: 490 orders (5.8% of total)
Top cities: Brooklyn (39), New York (38), Bronx (30), Staten Island (14), Binghamton (10)
PA: 311 orders (3.7% of total)
Top cities: Philadelphia (22), Pittsburgh (11), Butler (6), State College (6), West Chester (6)
Business Insights
1. Geographic Distribution:
- Cluster 1 (Eastern): Major metropolitan areas
- Cluster 2 (Central): Secondary markets
- Cluster 4 (Western): Growing markets
- Cluster 0 (Mountain): Emerging markets
- Cluster 3 (Pacific): Remote locations
2. Market Opportunities:
- Focus expansion in high-density states
- Target underserved cities in strong clusters
- Consider new distribution centers near cluster centers
3. Operational Recommendations:
- Optimize shipping routes within clusters
- Evaluate delivery times by cluster
- Consider regional pricing strategies
What-If Analysis Results
Shipping Duration Scenarios
| Scenario | Location | Description | Shipping Method | Price ($) | Predicted Duration |
|---|---|---|---|---|---|
| NYC - Holiday Season | (40.7128, -74.0060) | Urban delivery during holidays | EXPEDITED | $49.99 | 1.0 days |
| Rural Montana - Winter | (46.8797, -110.3626) | Rural delivery in winter | STANDARD | $29.99 | 1.0 days |
Order Volume Scenarios
| Scenario | Location | Description | Price ($) | Quantity | Predicted Volume |
|---|---|---|---|---|---|
| NYC - Holiday Season | (40.7128, -74.0060) | Urban delivery during holidays | $49.99 | 1 | 1.0 |
| Rural Montana - Winter | (46.8797, -110.3626) | Rural delivery in winter | $29.99 | 2 | 2.0 |
Average Rating Scenarios
| Scenario | Location | Description | Price ($) | MSRP ($) | Predicted Rating |
|---|---|---|---|---|---|
| NYC - Holiday Season | (40.7128, -74.0060) | Urban delivery during holidays | $49.99 | $59.99 | 4.6/5.0 |
| Rural Montana - Winter | (46.8797, -110.3626) | Rural delivery in winter | $29.99 | $34.99 | 4.6/5.0 |
Spatial Analysis System
Spatio-Temporal Data
- Spatial Data: Customer addresses, store locations, regional boundaries
- Temporal Data: Order dates, delivery dates, promotional periods
Visualization
- Heatmaps of sales intensity
- Choropleth maps by region
- Time-series plots of trends
Data Processing
Geocoding
Converting addresses into geographic coordinates using geopy
Spatial Operations
- Creating GeoDataFrame objects
- Spatial joins with external datasets
- Regional aggregation
Model Performance Analysis
Shipping Duration Model
Train R²: 0.9999
Test R²: 0.9999
Mean: 3.72 days
Std Dev: 2.47 days
Average Rating Model
Train R²: 0.9727
Test R²: 0.9697
Mean: 4.38/5.0
Std Dev: 0.75
Order Volume Model
Train R²: 0.9976
Test R²: 0.9978
Mean: 1.18 orders
Std Dev: 0.62
Features Used (27 total)
Geographic Features
- Latitude
- Longitude
- Distances to major cities
Shipping Features
- Shipping weight
- Shipping method
- Shipping tier
Temporal Features
- Order month/day
- Is weekend/holiday
- Seasonal indicators
Key Observations
- Shipping and Order Volume models show signs of overfitting
- Average Rating model shows good generalization
- Same feature set used across all models
- 5-fold cross-validation with 24 hyperparameter candidates
Recommendations
- Address overfitting through feature selection
- Implement regularization techniques
- Evaluate feature importance
- Check for potential data leakage
- Analyze prediction residuals
Technologies Used
Pandas
GeoPandas
Geopy
Seaborn
Statsmodels
Scikit-learn